ROS navigation stack is a set of packages and tools that enable you to build and run autonomous robots that can navigate in complex environments. It provides features such as localization, mapping, obstacle avoidance, path planning, and motion control. In this article, you will learn how to set up ROS navigation stack on your robot using a few basic steps.
Before you can use ROS navigation stack, you need to install ROS on your robot and your computer. ROS is an open-source framework that allows you to communicate and control your robot using various nodes, topics, and messages. You can choose from different distributions of ROS, such as Noetic, Melodic, or Kinetic, depending on your operating system and hardware compatibility. You can follow the official installation guide on the ROS website to install ROS on your robot and your computer.
Next, you need to configure your robot hardware to work with ROS navigation stack. You need to make sure that your robot has a base controller, a sensor for odometry, and a sensor for laser scan. These are the minimum requirements for ROS navigation stack to perform localization and mapping. You also need to create a URDF file that describes the physical structure and dimensions of your robot. A URDF file is an XML file that defines the links, joints, and frames of your robot. You can use tools like xacro or urdf_tutorial to create and visualize your URDF file.
For the ROS navigation stack to function, your robot hardware must be appropriately configured. Essential components include a base controller, an odometry sensor, and a laser scan sensor. These elements are critical for localization and mapping tasks. Additionally, you need to create a URDF file, which is an XML file that details your robot's physical structure, dimensions, and characteristics, including links, joints, and frames.
Building a robot needs lot of considerations -CAD modelling of the robot -Choosing the right motor with the torque applied over the wheels -Chassis - Robot simulation
Important side note: you can use a simulated robot in Gazebo. Turtlesim and Turtle bot in Gazebo would be a good starting point with pre-built examples and guides.
After you have configured your robot hardware, you need to launch the ROS nodes that are responsible for running ROS navigation stack. You can use roslaunch to launch multiple nodes with a single command. You need to launch at least four nodes: the base controller node, the odometry node, the laser scan node, and the move_base node. The base controller node communicates with your robot's motors and drives it according to the commands from the move_base node. The odometry node estimates the position and orientation of your robot based on the encoder readings from the wheels. The laser scan node publishes the data from your laser sensor as a LaserScan message on a topic. The move_base node is the core of ROS navigation stack. It integrates the localization, mapping, obstacle avoidance, path planning, and motion control functionalities. It subscribes to the odometry and laser scan topics and publishes the velocity commands to the base controller node.
You can make a modular launch file such that it has parameters that are set up launch and can also include/call other launch in a chronological order- think of this as a master launch file that does everything for you. To make life easier, you can also launch rqt parameter server and sims like Rviz with the world preloaded so you know what the sensors are currently observing.
Finally, you need to set up the navigation parameters that control the behavior and performance of ROS navigation stack. You can use YAML files to store and load the navigation parameters. You need to create at least three YAML files: the costmap_common_params.yaml, the local_costmap_params.yaml, and the global_costmap_params.yaml. The costmap_common_params.yaml file defines the common parameters for both the local and global costmaps. A costmap is a grid-based representation of the environment that assigns a cost value to each cell based on the occupancy and proximity of obstacles. The local_costmap_params.yaml file defines the parameters for the local costmap, which is a small window around the robot that is used for obstacle avoidance and local planning. The global_costmap_params.yaml file defines the parameters for the global costmap, which is a large map of the environment that is used for global planning and navigation goals. You can use the ROS navigation tutorials to learn how to set up the navigation parameters for your robot.
very important parameters of the navigation topics are -odometry values(x,y,z) -Transforms these are the aspects that play vital role in contribution of navigating a robot to and fro
Think clearly how your automation problem looks like and what problems have to be solved. Can it be simplyfied to 2D? Or does the problem space require 3D understanding? What motion model does your robot have? Is the given motion model common and part of Nav2?